• Applications

• Objectives

• Methods

• Interpretation

• aid in discovery of new populations of imperiled plants

• aid in creation of reserves under climate change models

• aid in predicting joint species distributions, i.e. obligate mutualisms

• fine scale models to study co-existence
• develop candidate species for metabarcoding

Besseya (=Synthris) alpina (A. Gray) Rydberg.

American Basin

B. alpina, Franklin #3948

• R

  • caret data partitioning Kuhn (2022)
  • sf vector data ala tidyverse Pebesma (2018)
  • terra raster data without headaches Hijmans (2022)
  • sdm modelling operations Naimi and Araujo (2016)

• grass gis many modules for creating predictors

• qgis graphical user interface for mouse guided visualization

• occurrences of a species in space (and time)

• Linear models:

  • check for spatial autocorrelation using Morans I
  • sampling artefact? remove samples
  • thin points piecewise by most offending record

• optional: pseudoabsences space (and time)
• more optional: absences space (and time)

• variables relating to patterns in biotic distribution
  • relevant to your extent and grains
  • require variation
  • no focus on factors governing biological processes, rather features which correlate with the known species distribution
• limitation: compute power

• domain: continental (e.g. North America)
  • maximum and minimum daily temperatures, precipitation amount and form, hydrologic drainage
• domain: regional (e.g. Southern Rockies)
  • elevation, soil classes, solar radiation
• domain: fine (e.g. East River Valley)
  • micro topography, water relations, shade, soils, species compositions

• ensembles pt. I & II
• linear models
  • assumptions - dependent variables (IV)
  • assumptions - independent variables (DV)
  • modelling
    • model evaluation
  • machine learning
    • assumptions ?
    • modelling
      • model evaluation
• ensembles pt. III

• problems with all models
  • garbage in -> garbage out
  • influential outliers
• with machine learning;
  • models can fixate on these observations
• solution:
  • run many models, synthesize the results

“we are stronger together than we are alone” - Walter Payton

• implementations:
  • weak learners; many simple decision tree models are combined to a single output
    • bootstrap aggregation (bagging) e.g. random forests
      • many models run independent of each
        
    • boosting e.g. boosted regression trees
      • many models run sequentially, focusing on correcting errors in the last iteration
  • stacking another form of combining many models
    • can accommodate the consensus output from bagged, boosted, or traditional linear models

• commonly implemented:
  • generalized linear models (glms)
  • generalized additive models (gams)

all evaluation performed by computer – too much information

much more common approach than individual linear models

species distributions are generally too complex for individual predictors, and building fully interactive terms would take a long time.

the typical approach since the late 90’s

do the work for you

none, get a few observations, the more the merrier.

train/test split (partition data)

• no free lunch
  • no silver bullet machine learning algorithm; each is able to work better than others under diverse circumstances.
  • try many types of models, select some that work for your application
• commonly implemented:
  • random forest
  • maximum entropy (maxent)

\[ Accuracy = \frac{\text{correct classifications}}{\text{all classifications }} \] \[ Sensitivity = \frac{\text{true positives}}{\text{true positives + false negatives }} \] probability of the method giving a positive result when the test subject is positive.

\[ Specificity = \frac{\text{true negatives}}{\text{true negatives + false positives }} \] probability of the method giving a negative result when the test subject is negative

• Area Under the ROC Curve (AUC)
  
• True Skill Statistic (TSS)
  • unequal sample sizes, i.e. evaluate many models at once
    
• Cohen’s kappa

• several r packages offer stacking of many models, based on your selection of evaluation criteria to weigh them
• the ‘test’ partition of your data are used to evaluate this final model

• any model, based on values present in gridded (raster) data can be predicted onto a new raster surface
• each covariate in the model, is generally a single raster layer
• r packages, such as terra, do all the work for you
• accordingly, species distribution models create a map as a product

• keep a lab notebook; experiments in silico ala experiments in vitro

• species distribution models are very simple!
• fun introduction to simple machine learning
• represent a hypothesis of the probability of suitable habitat
• new avenues (J-SDM’s) can include mutualisms
• stacked species (S-SDM’s) distributions for predicting ecological assemblies

• github/sagesteppe (preferred)
• reedbenkendorf2021@u.northwestern.edu

Ensemble learning utilizes many sets of trees, each tree being composed of many binary decisions, to create a single model. Each independent variable ( - or feature) may become a node on the tree - i.e. a location on the tree where a binary decision will move towards a predicted outcome. Each of the decision tree models which ensemble learning utilizes is a weak model, each of which may suffer due to high variance or bias, but which produce better outcomes than would be expected via chance. When ensembled these models generate a strong model, a model which should have more appropriately balanced variance and bias and predicts outcomes which are more strongly correlated with the expected values than the individual weak models.

Random Forest (RF) the training data are continually bootstrap re-sampled, in combination with random subsets of features, to create nodes which attempt to optimally predict a known outcome. A large number of trees are then aggregated, via the most common predictions, to generate a final classification prediction tree. Each individual prediction tree is generated independently of the others.

Boosted Regression Tree (BRT) (or Gradient Boosted tree) An initial tree is grown, and all other trees are derived sequentially from it, as each new tree is grown the errors in responses from the last tree are weighed more heavily so that the model focuses on selecting dependent variables which refine predictions. All response data and predictor variables are kept available to all trees.